A simple model of how language works, and one of the first developed, is as follows: we have a thought with a certain sense, we speak words forming a sentence with that same sense, the listener grasps the sense of our sentence and thus comes to have the same thought. This seems like an acceptable model, but it is lacking in some important details. What is a sense, how can words and thoughts have one, and how do they refer to objects? Frege, the originator of this terminology, gives an account of the sense as some kind of abstract object, but his explanation is somewhat unsatisfying, as how these senses interact with the physical world is somewhat mysterious. Russell provided a better explanation of what was going on by describing language as referring to objects through definite descriptions, which could be understood as the senses. Thus under such a model of language we have certain definite descriptions in mind, speak words expressing them, and then our words provide the listener with the necessary definite descriptions to think about the same things. But even this theory is inadequate when it come to describing how reference works because, as Donnellan pointed out, we can make reference to objects even if they don’t fit the definite description we are using to refer to them. His famous example of this phenomenon is the ability of two people to talk about “the man drinking the martini” even if he is really drinking something else, so long as it seems to the both of them that he is drinking a martini. From this Donnellan concludes that language really works in the following fashion: we have an intention towards some object, we speak some words, and through those words the listener is able to uncover our intention.
I don’t think that Donnellan’s analysis of how language works is complete, but it does introduce an interesting idea, namely that the words and structure of language are simply a tool for getting at each other’s thoughts, an inefficient way of mind reading if you will. And I think that Kripke’s causal theory about naming can’t account for this property of language, it leans too strongly on the idea that somehow certain words have the property of referring independently of people. But before I get to that I would like to fill out the model of language we have been developing here a little more. Specifically we need to deal with intentionality and how it can refer to objects. I think that this is where the definite descriptions fit into our story. Specifically the fact that our intentions can be said to objectively refer to things is that because they contain, consciously or unconsciously, some kind of description of that thing (Husserl called this the horizon). But when determining what someone’s intention references, in objective terms, we must relativize those intentions to the person (specifically to what the person knows and has experienced). If there are two identical sculptures, and the person only knows about one of them, then when that person thinks of the sculpture via some description of its appearance he or she is intending the one that they know of. Such a relativizing shouldn’t seem unusual, after all we seem to do it in our typical use of language. For example, if someone mentions the man who is drinking a martini we look for someone who looks like they are drinking a martini to them. If we have special knowledge that allows us to know that he really isn’t drinking a martini we will still grasp that the sentence we have heard refers to that man, since we know that to the other person the man seems like he is drinking a martini.
Finally, a quick note about names. Some wrongly assume that when we say that everything ultimately must be reduced to some kind of definite description that names of people are stand-ins for descriptions about how they look (or something similar). Kripke certainly seems to make this mistake. Really the description that underlies the name X is probably something like “the person who is called X”. Certainly such descriptions are allowable, because being called by a certain name, and recognizing that a name is ones own, is a property, as much as any social construct is a property. And in certain situations the description may be expanded to “the person called X by group Y” or the “the person called X who I stand in relationship Z with” when disambiguation is necessary. But I won’t go into such cases here.
Now let me turn to the causal theory of naming. The causal theory of naming says that a name refers to a specific object because at some point that object is dubbed with that name, and thereafter uses of the name refer to that object because they stand in the correct causal relation to the original naming event. There are, however, some examples that this theory of naming is simply unable to account for. Let us consider David and Goliath. Now it has come to my attention that the man David actually killed wasn’t Goliath, Goliath was killed by someone else. Given this, when we speak of Goliath who are we referring to, the man named Goliath by the Hebrews or the man David killed? We might think that the causal of theory of naming could be defended by appealing to the story, or our learning of the story, as a second naming, one which effectively gives the man David killed the name Goliath, at least for us. But the problem with this is that the story does refer to the “real” Goliath (as the giant, ect), it is just the ending that is mixed up. The causal theory of naming then is forced to pick one of these possibilities: either that we refer to the man David killed when we say Goliath, that we refer to the man the Hebrews called Goliath when we say Goliath, or that when we say Goliath we don’t properly refer to anybody. All three of these possibilities seem wrong. But the model of language I described above has no problem dealing with this odd situation. All we need to do in order to determine who the use of the name Goliath refers to is to determine who the speaker is intending. Is the speaker intending to talk about the man David killed or the man named Goliath by the Hebrews? Usually we can determine this by context and by a little knowledge about what the speaker knows (does the speaker know that David didn’t really kill Goliath?). Of course we might fail to grasp the speaker’s intentions, but all theories about language admit the possibility of miscommunication. So under the causal theory about naming this situation seems paradoxical, but under an intentional model of language it seems merely ambiguous, and I think this is good reason to prefer the intentional model.